Journal of Choice Modelling最新文献

This study presents the properties of the weibit travel choice model (which was newly developed in 2008) and its applications in transportation research. The weibit model uses a Weibull-distributed random error term, relaxing the identically distributed assumption to derive the closed-form choice probability expression. The properties and advantages of the weibit model are graphically illustrated by comparisons with the commonly used logit model. The relationship between the weibit and logit models are discussed in the transportation context. The applications of weibit models to transportation research are exemplified based on the mode choice modeling in various transportation networks incorporating emerging technologies. The study reveals that weibit models have attractive features for evaluating and planning future transportation systems.

Published choice experiments linked to various aspects of the COVID-19 pandemic are analysed in a rapid review. The aim is to (i) document the diversity of topics as well as their temporal and geographical patterns of emergence, (ii) compare various elements of design quality across different sectors of applied economics, and (iii) identify potential signs of convergent validity across findings of comparable experiments. Of the N = 43 published choice experiments during the first two years of the pandemic, the majority identifies with health applications (n = 30), followed by transport-related applications (n = 10). Nearly 100,000 people across the world responded to pandemic-related discrete choice surveys. Within health applications, while the dominant theme, up until June 2020, was lockdown relaxation and tracing measures, the focus shifted abruptly to vaccine preference since then. Geographical origins of the health surveys were not diverse. Nearly 50% of all health surveys were conducted in only three countries, namely US, China and The Netherlands. Health applications exhibited stronger pre-testing and larger sample sizes compared to transport applications. Limited signs of convergent validity were identifiable. Within some applications, issues of temporal instability as well as hypothetical bias attributable to social desirability, protest response or policy consequentiality seemed likely to have affected the findings. Nevertheless, very few of the experiments implemented measures of hypothetical bias mitigation and those were limited to health studies. Our main conclusion is that swift administration of pandemic-related choice experiments has overall resulted in certain degrees of compromise in study quality, but this has been more so the case in relation to transport topics than health topics.

This article introduces a basket and expenditure based choice experiment design to elicit consumer preferences for multiple products. This design is utilized to imitate a more realistic shopping scenario for consumers when choosing among many different products simultaneously. This approach allows participants to choose both multiple items, in this case vegetables, and related quantities/expenditures to place in a basket of goods. We provide an application of the experimental design to a vegetable choice experiment. This is done in conjunction with a sensory experiment to provide a contextual component to the experiment and econometric model. This type of experiment lends itself to the use of the Multiple Discrete-Continuous Extreme Value (MDCEV) class of models. More specifically, we use the extended version of the MDCEV model proposed by Palma and Hess (2020) that relaxes the need for a budget while also accounting for substitution and complementarity among products. We find that the proposed design and class of econometric methods present a flexible way to analyze consumer choice when the desire is to elicit preferences for a basket of goods rather than simple discrete alternatives or attributes.

Introduction

In order to solve the problems related to prior parameter misspecification in DCEs, Bliemer and Rose (2010) proposed a sequential approach where the design is updated after each respondent. This paper tries to find a more efficient alternative sequential method since the original proposal could be very time-consuming and expensive under some circumstances.

Methods

11 different strategies were simulated using 8 to 16 choice sets following a Monte Carlo approach. The accuracy and bias of the estimates of each strategy were studied using the relative error and mean value of their estimates.

Results

The DCE performs similarly to the original strategy by updating the design after five respondents. Among the other strategies, we discovered that, under certain circumstances, updating the design after 20 or 10 respondents led to accurate and not significantly biased estimates.

Conclusions

For a strategy to be efficient it might not be necessary to update the DCE after each respondent, but we found that updating the prior information relatively often and regularly can be almost as efficient as the original sequential proposal (for example, updating after five respondents might be a good choice). In addition, our findings suggest that each DCE has different efficient strategies depending on the number of attributes, levels, sets, and alternatives, so it can be concluded that a universal “optimal sequential strategy” does not exist.

Understanding the mechanism of continuous subscribing behavior is vital to the operation and even survival of customized bus (CB) systems. Capturing the differences in subscribing behavior of active and inactive users and developing targeted and differentiated marketing strategies accordingly will help to improve the user retention rate. The results of this study revealed differences in the subscribing behavior between active and inactive users and the influence of a wide range of factors on the subscription behavior between the two groups, such as area-related and individual characteristics. Travel distance presents an inverted “U” shape effect on the propensity to use CB with the vertex at around 16 km for active users and 28 km for inactive users, which is a further finding of the monotonic positive effect indicated by the existing studies. For every RMB 10 cheaper CB trip, the likelihood that an active user holds a higher “propensity” to use CB services will increase by 8%. The cost advantage will attract novelty-seeking travelers to try CB services occasionally, but cannot convert them into high-frequency users. The passengers who started to subscribe to the CB service in winter are more likely to travel by CB frequently no matter in active and inactive user groups. The findings of this empirical study lay a theoretical foundation for ridership retention and marketing orientation of the customized and demand-responsive transit services.

Brands develop strategies based on forecasts that allow for individual differences, usually attributed empirically to heterogeneity in consumers' preferences. Behavioral theories propose choice process heterogeneity as the conditioning stage for choice outcomes, and suggest that not accounting for it causes biases in parameters and policy measures. We conduct a Monte Carlo simulation to study how underlying choice process heterogeneity generates substantively significant biases in different market contexts if analysts (erroneously) channel heterogeneity solely into tastes. We extend the literature by using a game theoretical analysis, driven by the results from the demand simulation, to explore demand mis-specification effects on brands' profitability and market equilibrium. Through mixed strategies we examine necessary conditions for market equilibrium when managers have access to different demand representations but are uncertain about which is true. We demonstrate that biases generated by representing consumer response heterogeneity solely through preference heterogeneity are enough to significantly jeopardize brands' profits due to misalignment of firms' products and resources with demand. Our work forcefully demonstrates to both marketers and econometricians/data scientists the necessity of modeling choice process heterogeneity given its impacts on brands’ performance.

We propose a new modelling approach for daily activity scheduling which integrates the different daily scheduling choice dimensions (activity participation, location, schedule, duration and transportation mode) into a single optimisation problem. The fundamental behavioural principle behind our approach is that individuals schedule their day to maximise their overall derived utility from the activities they complete, according to their individual needs, constraints, and preferences. By combining multiple choices into a single optimisation problem, our framework is able to capture the complex trade-offs between scheduling decisions for multiple activities. These trade-offs could include how spending longer in one activity will reduce the time-availability for other activities or how the order of activities determines the travel-times. The implemented framework takes as input a set of considered activities, with associated locations and travel modes, and uses these to produce empirical distributions of individual schedules from which different daily schedules can be drawn. The model is illustrated using historic trip diary data from the Swiss Mobility and Transport Microcensus. The results demonstrate the ability of the proposed framework to generate complex and realistic distributions of starting time and duration for different activities within the tight time constraints. The generated schedules are then compared to the aggregate distributions from the historical data to demonstrate the feasibility and flexibility of our approach.

Latent class (LC) models have been used for decades. In some cases, models of this kind have exhibited difficulties in identifying distinct classes. Identifiability is key to determining the presence or absence of the different population cohorts represented by the latent classes. Theoretical identifiability addresses this issue in general, but no empirical identifiability analysis of this kind of model has been performed previously. Here, we analyse the theoretical properties of LC models to establish necessary conditions on the classes to be identifiable jointly. We then, establish a measure of behavioural difference and relate it to empirical identifiability; this measure highlights factors that are crucial for identifiability. We show how these factors affect identifiability through simulation experiments in which classes are known, and test elements such as the proportion of individuals belonging to each latent class, different correlation structures and sample sizes. In our experiments, each latent class corresponds to a different choice heuristic. We present a graphical diagnostic that supports the measure of behavioural difference that promotes identifiability and provide examples of model non-identifiability, partial identifiability, and strong identifiability. We conclude by discussing how non-identifiability can be detected and understood in ways that will inform survey design and analysis.